When the objective is to stack separate result sets , the Analysis Editor supports set operations such as UNION, UNION ALL, INTERSECT, and DIFFERENCE . Oracle documents a specific Combine Results function that lets authors create criteria from another subject area and apply a set operator to the two query result sets.

Set operations, however, aren't relational join definitions. They combine compatible result sets and require matching numbers of columns and compatible data types. They don't create primary-key/foreign-key or other semantic relationships between subject areas.

Cross-subject-area relational behavior depends on the semantic model , including common dimensions, facts, and relationships defined for the participating subject areas. Oracle also supports cross-subject-area analyses using conformed dimensions, but those relationships derive from model metadata rather than a join created ad hoc through set-operation syntax.

Dashboard prompts simply provide runtime filtering and don't create data relationships. Likewise, adding a second query doesn't itself define semantic join logic.

Therefore, the accurate distinction is: use set operators when stacking compatible result sets; define and govern actual joins through the semantic model .

Reference Topic: Managing and Creating Analyses, Views, and Content — Combining Subject Areas and Set Operations

===============